AgriTech Design Lab

Lab Overview

The Agri-AI Design Lab is the capstone experience of the Fataplus Bootcamp, where participants ship a production AgriTech product for real farmers in Madagascar. The Smart Irrigation Companion serves as the foundational case study, combining UX research, AI strategy, no-code development, and field deployment.

Live Project: Smart Irrigation Companion

Client: FOFIFA Agritech Lab
Users: 800 rice farmers in Alaotra-Mangoro region
Timeline: 14-week pilot (November 2025 - February 2026)
Budget: €35,000
Goal: Deploy bilingual mobile/web app with predictive watering alerts, cooperative scheduling, and impact analytics

Design Sprint Case Study

Problem Statement

Context: Smallholder farmers lack timely irrigation guidance, causing water waste and yield loss. User Pain Points:

Farmers copy neighboring plots instead of using scientific timing
Cooperative schedulers manage 65+ farmers with paper logs and WhatsApp
Unreliable weather forecasts and no pump availability visibility
Manual scheduling causes conflicts and inequitable water distribution

Success Criteria:

15% yield increase
20% water savings
70% weekly active users during pilot
NPS ≥35
Onboarding completion under 15 minutes

Engagement Kickoff

The project launched with a structured kickoff canvas defining stakeholders, constraints, and workstreams:

Stakeholder Map

Stakeholder	Role	Expectations	Communication
Dr. Ranaivo	Program Lead (FOFIFA)	Milestone visibility, impact metrics	Weekly email summary + monthly review call
Lova	Cooperative Champion	Training materials, offline access	WhatsApp group; on-site sessions bi-weekly
FOFIFA Data Team	Data Providers	Clear API specs, governance alignment	Slack connect; data clinic Fridays
Fataplus Design Squad	Delivery Team	Sprint scope alignment, support	Daily async standup in Linear
Bootcamp Cohort	Trainee Testers	Structured exercises, mentor feedback	Weekly studio lab

Constraints & Assumptions

Connectivity: Spotty 3G coverage requiring offline-first design
Devices: 68% use basic Android phones (≤30MB app size limit)
Infrastructure: Limited solar pumps, shared cooperative equipment
Budget: €35k covering design, development, field deployment, and training
Timeline: 14 weeks from kickoff to impact readout
Compliance: Align with local data privacy norms and MAEP reporting standards

Workstream Ownership

Agent / Team	Responsibilities	Immediate Next Step
Studio Orchestrator	Coordination, playbook exports	Schedule sprint zero + confirm tool stack
Field Ethnographer	Research synthesis, persona updates	Digest 12 interview transcripts by Nov 14
AI Solution Crafter	Use-case prioritization, data audit	Map AI opportunities + launch readiness audit
Product Experience Engineer	UX/UI + no-code build	Draft navigation concept and component inventory
CX/DX Strategist	Adoption & KPI framework	Outline training path + KPI dashboard blueprint
Bootcamp Mentor	Education alignment	Adapt sprint into cohort exercise brief

Research & Personas

Field Research Synthesis

The research digest compiled insights from farmer interviews, cooperative training sessions, and API documentation:

Source	Type	Key Insight	Evidence	Follow-up
Interview_RiceFarmer_001	Transcript	Farmers rely on neighbors for irrigation timing	”Nous attendons que le champ d’à côté commence”	Capture irrigation diary over 4 weeks
CoopTraining_Sept2025	Notes	Need offline-first onboarding	Trainers request printable guides	Prototype SMS onboarding script
MeteoMada_API	API spec	Hourly forecast granularity available	API documentation	Confirm pricing & rate limits
PilotSurvey_July2025	CSV	68% use basic Android phones	Survey data	Define minimum device requirements

Research Gaps Identified:

Missing evapotranspiration historical data for 2023
2 Malagasy audio interviews require transcription
Need irrigation ritual shadowing sessions
Compile agronomist heuristics for AI prompt design

Persona Spotlight: Voahirana Randria

The primary persona represents cooperative irrigation schedulers balancing farmer equity with limited resources:

Voahirana Randria - Cooperative Irrigation Scheduler

Location: Andilamena cooperative, Alaotra-Mangoro
Responsibility: Coordinates watering schedules for 65 rice farmersStory:
Voahirana juggles paper logs, WhatsApp messages, and weekly meetings to prevent crop stress while ensuring equitable pump access. She reports water usage to MAEP and donors but lacks visibility into pump downtime and forecast reliability.Primary Job-to-be-Done:
Ensure equitable irrigation slots so every farmer receives water at the optimal timeSecondary JTBD:
Report water usage and crop status to MAEP and donors with proof of impact

Pains & Gains

Pains	Gains
Manual scheduling causes conflicts	Wants predictive guidance
Forecasts unreliable	Automated alerts
Lacks visibility into pump downtime	Proof of impact for donors
Paper logs difficult to compile for reports	Dashboard with SDG alignment

Tech & Channel Preferences

Devices: Android phone with intermittent 3G; cooperative laptop shared once weeklyTrust Anchors: FOFIFA agronomists, cooperative eldersCommunication Channels: WhatsApp for coordination, SMS for alerts, phone calls for urgent issuesLiteracy: Fluent in Malagasy and French; prefers bilingual interfaces

AI/UX Opportunities

Dashboard highlighting upcoming stress periods with weather context
SMS alerts for schedule changes and pump conflicts
Offline-ready forms for logging irrigation sessions
Auto-generated reports for MAEP with bilingual summaries
Alternative slot suggestions when conflicts detected

Validation Partners: Lova (cooperative champion), FOFIFA agronomists, MAEP regional office

Concept Development

AI Concept Portfolio

The team prioritized three core concepts using impact/feasibility matrices:

1. Predictive Watering Advisor

Impact: High
Feasibility: MediumSMS and app alerts with explainable recommendations combining weather forecast, crop stage, and pump availability

2. Cooperative Scheduling Optimizer

Impact: High
Feasibility: HighDrag/drop calendar resolving pump conflicts with automated alternative slot suggestions

3. Impact Analytics Coach

Impact: Medium
Feasibility: HighAutomated reporting to MAEP and donors with SDG tracking and yield comparisons

Data Readiness Audit

Before prototyping, the team assessed AI feasibility through a comprehensive data audit: Overall Rating: Medium-Low Available Data Sources:

MeteoMada API (hourly weather forecasts)
Pump usage logs (currently paper-based)
Farmer plot records (cooperative spreadsheets)
Survey data from pilot participants

Data Gaps & Remediation:

Gap	Impact	Remediation	Owner	Timeline
Pump logs not digitized	High	Create Supabase forms for cooperative	Lanto	Week 2
No evapotranspiration data	Medium	Secure MeteoMada API contract with caching	FOFIFA Data Team	Late November
Missing data stewards	Medium	Assign cooperative coordinators	Lova	Week 1
Privacy framework undefined	High	Implement anonymization in Supabase	Fataplus + Legal	Week 3

UX/UI Design

Experience Architecture

The product supports three primary journeys:

Farmer Alert Journey

Farmer receives predictive watering alert → confirms readiness → logs outcome via mobile app or SMS

Cooperative Scheduling

Cooperative dashboard for viewing weekly schedule → resolving conflicts → notifying affected farmers

Agronomist Insight Review

Review AI recommendations → approve or override → provide context for edge cases

Edge Cases & Offline Considerations:

SMS fallback for low connectivity
Queue notifications when pump unavailable
Manual override logging with agronomist approval
Voice prompts for farmers with literacy constraints

Key Screens & Components

Mobile App (FlutterFlow)

Alert Feed:

Card-based list of irrigation recommendations
Weather summary with visual icons
Confirmation button with SMS fallback
Offline badge when cached

Irrigation Schedule Timeline:

Weekly calendar view with farmer assignments
Pump availability indicators
Conflict warnings with alternative slots
Drag-to-reschedule (cooperative coordinators only)

Weather Insights Card:

7-day forecast with rain probability
Evapotranspiration rates
Optimal irrigation windows highlighted

Onboarding Wizard:

3-step setup: Profile → Plot details → Notification preferences
Bilingual FR/MG with audio prompts
Skip option for SMS-only mode

Web Dashboard (Bubble)

Cooperative Admin Dashboard:

Weekly schedule calendar with drag/drop
Conflict detection with automated resolution suggestions
Farmer contact list with alert history
Pump maintenance log

Analytics & Reporting:

KPI summary cards (yield lift, water savings, active users)
Weekly usage charts by farmer and plot
Export to Google Sheets for MAEP reporting
SDG alignment indicators

Agronomist Review Panel:

AI recommendation queue
Approve/override interface with context fields
Feedback log from farmers
Model performance metrics

Design System Tokens

Typography:

Bilingual font stack (Noto Sans for Latin/Malagasy)
Size scale: 12px (caption) → 48px (hero)
Line height 1.5 for readability in sunlight

Color Palette:

High contrast for outdoor use (WCAG AAA)
Primary: Agriculture green (#2D7A3E)
Alert: Warning amber (#F59E0B)
Success: Growth teal (#10B981)

Components:

Buttons (primary, secondary, ghost, SMS-link)
Cards (elevated, outlined, interactive)
Status badges (online, offline, syncing, conflict)
Toast notifications (success, error, info)

Customer Journey Map

The full lifecycle journey for Voahirana (cooperative scheduler):

Stage	User Goal	Touchpoints	Emotions	Metrics	Opportunity
Awareness	Understand value of smart irrigation	Cooperative briefing, poster, SMS teaser	Curious yet skeptical	Event attendance rate	Use success stories from pilot farmers
Consideration	Evaluate feasibility & effort	Demo day, one-on-one with Fataplus	Hopeful but cautious about workload	Sign-up conversions	Offer quick-start kit + offline brochure
Onboarding	Configure schedules & alerts	Mobile wizard, WhatsApp prompts	Energized but overwhelmed by setup	Onboarding completion time	Guided setup sessions, interactive checklist
Activation	Receive and act on first alerts	Push notification, SMS, phone check-in	Confident if pump ready, anxious if conflicts	Alert acknowledgment rate	Provide automated alternative slots + hotline
Adoption	Integrate into weekly routine	Dashboard reviews, cooperative sync	Empowered when data aligns, frustrated by gaps	Weekly active usage	Simplify dashboards, add printable summaries
Impact Reporting	Share results with sponsors	Analytics dashboard, MAEP report	Proud if metrics positive	Impact report submissions	Auto-generate bilingual reports
Advocacy	Recommend to other cooperatives	Showcase events, bootcamp shareback	Motivated, community pride	Referral count	Provide referral incentives, highlight stories

Improvement Experiments:

Test voice-note instructions for low-literacy farmers
Pilot community leaderboard to gamify water savings
Align alerts with local radio bulletins for redundancy

No-Code Development

Build Strategy

Technology Stack

Frontend:

Bubble (web admin for cooperatives and agronomists)
FlutterFlow (mobile companion for farmers)

Backend:

Supabase (PostgreSQL database with real-time sync)
Twilio (SMS integration for offline fallback)

Integrations:

MeteoMada API (weather forecasts)
Google Sheets (archival and MAEP reporting)
Zapier (escalation workflows)
OpenAI GPT-4 (contextual irrigation tips)

Data Model

Core Entities:

Farmer
  - id, name, phone, language_preference, cooperative_id
  
Plot
  - id, farmer_id, crop_type, area_hectares, soil_type
  
Pump
  - id, cooperative_id, capacity_lpm, status, maintenance_log
  
ScheduleSlot
  - id, plot_id, pump_id, start_time, duration_minutes, status
  
Alert
  - id, farmer_id, recommendation, weather_context, sent_at, acknowledged_at
  
FeedbackEntry
  - id, farmer_id, alert_id, success_rating, notes, agronomist_review

Sprint Schedule

Sprint	Focus	Deliverables	Owner	Validation
0 - Setup	Platform setup, design tokens	Bubble workspace, FlutterFlow project, Supabase schema draft	Lanto	Internal review
1 - Alerts MVP	Alert feed, SMS workflow	Alert screen, Twilio integration, dashboard skeleton	Lanto + Hery	Farmer pilot with 5 users
2 - Scheduling	Cooperative calendar	Drag/drop schedule, conflict resolution rules, email notifications	Lanto	Cooperative simulation workshop
3 - Analytics	Impact tracking, training	Dashboard with KPIs, training materials, export function	Rado	Field test + KPI baseline

Automation Workflows

Bubble Workflows

Daily Alert Queue (Runs at 06:00):

Query MeteoMada API for weather forecast
Calculate irrigation recommendations per plot
Check pump availability and detect conflicts
Queue alerts for farmers via Supabase

Schedule Update Notification:

Trigger: ScheduleSlot modified
Action: Send SMS to affected farmer with new time
Log notification in Alert table

Conflict Escalation:

Trigger: Conflict detected >24h unresolved
Action: Zapier webhook to notify cooperative coordinator
Create task in Linear for Fataplus support

AI Prompt Kits

Irrigation Recommendation Explainer:

Prompt: "Explain irrigation recommendation"
Context:
  - Weather forecast (rain probability, temperature, evapotranspiration)
  - Crop stage (growth phase, days since planting)
  - Soil moisture (if sensor available, otherwise estimated)
Output:
  - Bilingual SMS (FR/MG) with actionable timing
  - Rationale: "Rain expected tomorrow, delay by 24h"
Model: GPT-4
Temperature: 0.3 (for consistency)
Guardrails:
  - Never recommend irrigation during forecast rain >70%
  - Flag if soil moisture sensors offline
  - Defer to agronomist if temperature anomaly detected

Weekly Water Usage Summarizer:

Prompt: "Summarize weekly water usage"
Context:
  - Farmer feedback logs
  - Pump usage data (start/end times, flow rates)
  - Yield estimates (if available)
Output:
  - Dashboard card with savings percentage
  - SDG alignment indicators (SDG 6: Clean Water)
  - Comparison to baseline week
Model: Claude 3.5 Sonnet
Temperature: 0.5 (for narrative variety)

Offline-First Caching

FlutterFlow Offline Plugin:

Cache last 3 alerts locally
Store farmer profile and plot details
Queue outgoing feedback submissions
Sync when connectivity restored

SMS Fallback Logic:

If push notification fails → send SMS within 5 minutes
SMS includes: Recommendation + Weather summary + Confirmation keyword
Farmer replies with keyword → update acknowledgment in Supabase

Sync Indicators:

Badge showing “Offline” or “Last synced: 2h ago”
Manual refresh button
Background sync every 30 minutes when online

Field Deployment

Training & Onboarding

Delivery Enablement Materials:

Video Tutorials (FR/MG with subtitles): App walkthrough, SMS keyword guide, dashboard navigation
Printable Quick-Start Guides: 1-page visual guides for farmers and cooperative coordinators
WhatsApp Micro-Lessons: Daily tips sent to cooperative group chat
Radio Bulletin Scripts: Align alerts with local radio for redundancy
Bootcamp Lab Integration: Use engagement as teaching case study for 2026 cohort

Training Workshops:

Co-Design Session with Voahirana

Test scheduling flows and offline-first prototypes with cooperative peers

Farmer Onboarding Sessions

Guided setup of mobile app with mentor support, SMS fallback configuration

Agronomist Training

AI recommendation review panel, override workflows, model performance interpretation

MAEP Reporting Tutorial

Dashboard walkthrough, export functions, SDG alignment metrics

Validation Checkpoints

Week	Checkpoint	Participants	Deliverables	Success Criteria
2	Concept Review	FOFIFA leadership, Fataplus team	Concept portfolio, data audit	Approve top 3 concepts, data remediation plan
6	Prototype Field Test	5 pilot farmers, Lova (champion)	FlutterFlow MVP, SMS workflow	≥4/5 satisfaction, identify 3 priority fixes
12	MVP Launch	65 farmers, cooperative coordinators	Full app + dashboard, training materials	70% onboarding completion, 50% weekly active
14	Impact Readout	All stakeholders, bootcamp cohort	Impact report, MAEP submission	Yield +10%, water savings +15%, NPS ≥30

Bootcamp Integration

Lesson Kit: Smart Irrigation Design Sprint

The capstone project was adapted into a structured bootcamp lesson for the 2026 cohort: Session Date: November 18, 2025
Cohort: Figma EDU Bootcamp 2026 – Pilot Lab Objectives:

Translate field research into AI-enabled product concepts
Craft UX flows and no-code briefs aligned with agritech constraints
Practice bilingual communication and inclusive design for farmers

Exercises:

Insight-to-Concept Mapping: Map top 3 insights to AI concepts; justify with impact/feasibility matrix
Alert-to-Action Micro-Journey: Design flow with offline fallback screens
Optional Extension: Prototype MVP component in FlutterFlow using provided data schema

Resources Provided:

Research digest (research-digest-smart-irrigation-2025-11-10.md)
Persona profile (Voahirana Randria)
Concept-to-prototype plan
No-code sprint template
Journey map canvas

Cohort Feedback (November 19, 2025)

Team Performance Summary:

Team	Highlights	Opportunities	Next Steps
Team Alaotra	Strong persona empathy, clear SMS flows	Need deeper API feasibility analysis	Pair with AI Solution Crafter to refine data audit
Team Betsiboka	Innovative cooperative dashboard, solid Figma tokens	Overlooked offline fallback in FlutterFlow	Implement low-bandwidth mode and retest
Team Canal+	Great storytelling tying SDG metrics	Concept drifted from farmer needs mid-sprint	Revisit insight matrix, align features with primary JTBD

Overall Assessment: Cohort teams successfully translated agritech insights into actionable AI concepts and UX prototypes. Most groups delivered bilingual narratives and considered offline-first constraints. Next sprint focus: Refining data readiness assumptions and testing no-code performance. Upcoming Milestones:

Nov 22: Submit revised prototypes with offline mode adjustments
Nov 25: Joint review with cooperative champions
Dec 2: Final bootcamp expo rehearsal

Impact & Metrics

Target Outcomes

Yield Lift

Target: +15%
Measured by cooperative harvest records compared to baseline

Water Savings

Target: +20%
Tracked via pump usage logs and farmer feedback

User Adoption

Target: 70% weekly active
Alert acknowledgment and dashboard login rates

Satisfaction

Target: NPS ≥35
Post-pilot survey with farmers and cooperative coordinators

Analytics Instrumentation

Supabase Event Tracking:

Alert sent → acknowledged → irrigation executed
Schedule conflict detected → alternative suggested → resolved
Feedback submitted → agronomist reviewed → action taken
Dashboard visited → report exported → MAEP submitted

Weekly Automated Reports:

Water usage per farmer and plot
Alert acknowledgment rates
Schedule conflict resolution time
Offline mode usage frequency
Farmer feedback sentiment analysis (AI-powered)

Real-World Impact (Projected)

For 800 rice farmers in Alaotra-Mangoro:

Economic: Estimated €120k additional revenue from 15% yield increase
Environmental: 1.6M liters water saved annually (20% reduction)
Social: Equitable pump access reducing cooperative conflicts
SDG Alignment: SDG 2 (Zero Hunger), SDG 6 (Clean Water), SDG 13 (Climate Action)

For Bootcamp Participants:

Portfolio: Production app with real users for job applications
Skills: End-to-end product design from research to deployment
Network: Connections with FOFIFA, cooperatives, and AgriTech ecosystem
Certification: Validated through live project impact metrics

Playbook & Workflows

The entire engagement was orchestrated using BMAD agent workflows:

Core Workflow Sequence

Engagement Kickoff: *kickoff-engagement (Studio Orchestrator)
Asset Ingestion: *ingest-assets (Field Ethnographer)
Insight to Concept: *map-ai-usecases (AI Solution Crafter)
Concept to Prototype: *concept-to-prototype (Product Experience Engineer)
Data Readiness Audit: *data-readiness (AI Solution Crafter)
Service Blueprint: *plan-adoption (CX/DX Strategist)
CX Journey Map: *cx-journey (CX/DX Strategist)
Bootcamp Lesson Kit: *create-lesson (Bootcamp Mentor)
Playbook Export: *export-playbook (Studio Orchestrator)

Agent Quick Commands

Agent	Trigger	Description
Studio Orchestrator	`kickoff-engagement`, `export-playbook`	Run orchestration workflows
Field Ethnographer	`ingest-assets`, `persona-forge`	Research synthesis
AI Solution Crafter	`map-ai-usecases`, `data-readiness`	AI opportunity analysis
Product Experience Engineer	`concept-to-prototype`, `nocode-sprint`	UX + no-code delivery
CX/DX Strategist	`plan-adoption`, `cx-journey`	Adoption planning
Bootcamp Mentor	`bootcamp-mode`, `create-lesson`	Training workflows

Validation Checklist

Next Steps

UX/UI Program

Explore the full curriculum and module structure

Bootcamp Overview

Return to program overview and certification

Project Partners: FOFIFA, MeteoMada, Alaotra-Mangoro Cooperatives
Bootcamp Cohort: Figma EDU Bootcamp 2026 – Pilot Lab
Funding: PIC Pole Intégré de Croissance Madagascar

Get Started

Platform Features

Design System

Bootcamp

AgriTech

Deployment

​Lab Overview

Live Project: Smart Irrigation Companion

​Design Sprint Case Study

​Problem Statement

​Engagement Kickoff

​Research & Personas

​Field Research Synthesis

​Persona Spotlight: Voahirana Randria

Voahirana Randria - Cooperative Irrigation Scheduler

​Concept Development

​AI Concept Portfolio

1. Predictive Watering Advisor

2. Cooperative Scheduling Optimizer

3. Impact Analytics Coach

​Data Readiness Audit

​UX/UI Design

​Experience Architecture

​Key Screens & Components

​Customer Journey Map

​No-Code Development

​Build Strategy